pensions core course 2013: expanding coverage - recent developments

EXPANDING COVERAGE: RECENT

DEVELOPMENTS

Robert Palacios, World Bank

Pension Core Course

April 2013

The policy challenge 2

LICs – 17% MICs – 51% TSEs – 66%

The policy challenge 3

Decades of stagnant coverage – failure to expand

traditional SI, payroll tax-financed model

Falling coverage in Eastern Europe and FSU

Increasing concern as schemes mature and require

high payroll tax rates that may further distort the

labor market

Aging populations that will put pressure on both

health and pensions (IMF 2011, 2012).

Emerging institutional responses 4

ILO’s Social Floor (UN, DFID)

Santiago Levy’s analysis and proposal

World Bank

social pensions and MDCs covered in SP strategy

MDC book

Social insurance and labor markets book

upcoming LAC book

upcoming ECA flagship report on pensions

Emerging responses on the ground 5

Expansion of social pensions

Chile, India, Nepal, Mexico, Panama

New experiments with matching contributions

China, India, Turkey

Subsidized health insurance premia for the poor

Colombia, Georgia, Ghana, India, Pakistan, Turkey

Remains to be seen what happens as ECA country

pension coverage begins to decline

6

Country

types

PPP$YCAP Coverage ratio Ratio 20-59/60+

population

LIC >4500 17% 7.6

MIC 4500-15,000 51% 6.3

HIC 15,000+ 90% 3.4

TSE 2000-20,000 66% 3.7

7

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

1 2 3 4 5

Quintile

% o

f e

mp

loy

ed

co

ntr

ibu

tin

g

Uruguay

El Salvador

India

Characteristics of target population 8

Variable income flows

Low savings capacity

Low exposure to formal financial sector

Transient career path (rural-urban migrants)

High degree of self-insurance (i.e., lack of various types of insurance coverage)

High discount rate/liquidity preference

Higher mortality/morbidity (relative to covered)

Few opportunities for formal jobs (evasion issue minor)

Strategies for system design 9

Minimize transaction costs

Allow small and variable contribution amounts and flexible timing

Harness existing groups where possible

Use IT to lower transaction costs on front end (banking correspondents, mobile payments)

Use formal pension system infrastructure where feasible

Simple investment types, reliance on defaults

Effective outreach

Credible institutions must participate on provider side

Pull factor may require paying providers’ incentives for enrolment (especially at outset)

Strategies for system design 10

Affordability and incentives

Affordable contribution levels

Link with health/disability insurance where feasible

Voluntary pensions in rich countries exist due to tax

treatment, but irrelevant for informal sector workers – a

substantial matching contribution is needed to overcome

high discount rate and liquidity preference

Age of withdrawal must be in line with realistic

biological deterioration

Matching contributions 11

MCs can be targeted and capped (e.g., India matches up

to 1000 rupees per year for informal sector workers)

1:1 match provides an immediate 100% return

Can be bundled with insurance (eg. Health) reducing

the need for liquid precautionary savings

In practice, there are a number of countries planning or

starting MDCs, but not enough experience to have

robust empirical results

China, India, Indonesia, Vietnam, Dominican Republic

Steps in setting key parameters for MDC

MCT

TBL = x % of MCL

CR/PL = $

AP = % of PL

SR = 1 - AP

MCT – minimum consumption target

TBL – target benefit level

CR – contribution rate

PL – premium level

AP – ability to pay

SR – subsidy rate

CT – coverage target

BR – budget requirement

CT - # workers BR - $

How affordable is the MDC to workers?

13

Example: Based on a target benefit just above the Indian poverty line, the contribution required with a 1:1 match is 5% of income for decile 3

0%

2%

4%

6%

8%

10%

12%

3 4 5 6 7 8

decile

% o

f in

co

me

How much take up can you buy?

14

Duflo et. al. (2005) tested the take up elasticity for US low income workers, but similar studies have not been done for developing countries

0

5

10

15

20

25

30

35

40

45

50

0 50 100 150 200 250

% t

ak

e u

p

% matching contribution

How much take up can you buy? 15

Age should be linked to target group 16

Some limited empirical evidence supports the

intuition that the life expectancy differential by

income level is greater in LICs (eg., Bannerjee and

Duflo (2005) and Pal and Palacios (2010))

Age of withdrawal should take shorter life

expectancy of target population into account

Age can also be coordinated with social pension

eligibility age

Age parameters 17

10

12

14

16

18

20

22

0 5000 10000 15000 20000 25000 30000 35000 40000 45000

PPP adjusted $ income per capita

life

exp

ecta

ncy a

t ag

e 6

0

Parameters must be consistent with fisc

18

Rough estimate of MDC cost

Set target pension at 40% YCAP

Calculate required total contribution for full career 10% of

YCAP

Set match at 1:1 (5% of YCAP from government)

With labor force/population (40%)* share in informal sector

(80%) * take up (50%) = 0.8% of GDP

This can be reduced to 0.4% of GDP if targeted to the

bottom half of the informal sector; (this yields a 40%

increase in coverage or 8 percentage points)

Match can be reduced subject to fiscal constraints

MDCs and social pensions

19

MDCs take a long time to mature and have no impact on old age poverty; it does nothing for the current or soon to be old

MDC policy and social pensions can be linked and harmonized to achieve clear objectives over time

Social pension dependence will be greater for older workers and gradually be replaced by dependence of younger workers on MDCs

SP can be set at absolute poverty level and indexed to inflation while MDC parameters linked to YCAP; prefunding as population ages

Set target pension at 40% YCAP

Calculate required total contribution for full career 10% of YCAP

Set match at 1:1 (5% of YCAP from government)

With labor force/population (40%)* share in informal sector (80%) * take up (50%) = 0.8% of GDP

This can be reduced to 0.4% of GDP if targeted to the bottom half of the informal sector; in this case, coverage would be doubled

Match can be reduced subject to fiscal constraints

Social pension only

20

0

.25

.5

.75

1

1.25

Gro

ss r

ep

lace

me

nt ra

te

0 .5 1 1.5 2 2.5 3

Individual income, proportion of average income per capita

Contributory Social

0

.5

1

1.5

2

2.5

Gro

ss r

ela

tive p

en

sio

n leve

l

0 .5 1 1.5 2 2.5 3

Individual income, proportion of average income per capita

Contributory Social

Here there is complete reliance on a social pension indexed to income with no MDC

Social pension and MDC

21

0

.5

1

1.5

2

2.5G

ross r

ela

tive p

en

sio

n leve

l

0 .5 1 1.5 2 2.5 3

Individual income, proportion of average income per capita

Contributory Social

Here, the social pension indexed to prices, with MDC parameters indexed to income

0

.25

.5

.75

1

1.25

Gro

ss r

ep

lace

me

nt ra

te

0 .5 1 1.5 2 2.5 3

Individual income, proportion of average income per capita

Contributory Social

22

0%

2%

4%

6%

8%

10%

12%

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61

% o

f G

DP

YEARS

Costs

MDC w/SP

SP only

Concluding thoughts

23

MDCs are being considered or started in a number of countries with low coverage

MDC policies alone do not address the current coverage gap – social pensions can play that role in the short run until MDC matures

Careful analysis of fiscal tradeoffs between the two types of program can only be done with long term projections and studies of take up elasticity

It may be especially attractive in countries with DC schemes for formal sector workers to reduce start up costs and allow for a seamless system